poster - Boulder County

Counter Accuracy vs. Distance from Counter
100.00%
90.00%
4
TB3
50.00%
40.00%
30.00%
ART4
(WB-40)
20.00%
10.00%
0.00%
0
10
20
30
40
50
Vehicle Speed (mph)
60
70
80
This graph plots the wheelbase and speed of
each vehicle in the motorcycle class from a
typical 24-hour traffic count. Our equipment was
counting some bicyclists, though the accuracy
was unknown.
0.00%
1
2
3
4
5
6
7
8
BOCO Rules for T4:
BOCO Rules
for 2Bikes:
4 or more axles
2 axle groups
4 or more axles
2 axle groups
d1 > 4 ft
4 axles
2 axle groups
2.9 ft < d1 < 4 ft
2.9 ft < d3 < 4 ft
Bin Number
Traffic counters use classification
schemes to sort vehicles by type,
or class. We determined groups of
bicyclists were being misclassified by
the ARX Cycle classification scheme
the counters were using.
ART6
non standard
(!
!
(
(
!
!
(
!
(
!
# NEVA RD
*
(
!
(
(!
!
7 Day Counts- 2 Counters
#NIWOT RD (
*
!
#
*
1000
1000
1000
! 24 Hour Counts- 2 Counters
(
!
± 27
± 99
± 90
Unpaved Road- No Bike Counts
R
!
(
(
( !
!
!
(
!
(!
(
(
(!
!
!
(
(
!
(
!
DE
ND
R
(
!
#
*
(
!
!
(
(
!
! JAY RD
(
( !
!
(
(
!
(
!
(
!
(
!
(
! !
(
(
!
(
!
T RD
MON
VAL
(
!!
( (
!
(
!
(
!
VALMONT RD
!
(
! !(
(
(
!
(
!
(
!
!!
(
(
(
!
(
!
!
(
(
!
(
!
DR
(
!
(!
!
(
!
(
(
!
!
(
(
!
(
!
(
!
(
!
BASELINE RD
LOOKOUT RD
(
!
(
!
!
(
(
!
!
(
(
( !
!
(
!
(
!
(
!
(
!
!!
(
(
(
!
(
!
(
!
(
!
#
*
(
!
NIWOT RD
!!
(
(
( !
!
(
(
!
#
*
LIN
(
!
(
!
(
!
(
!
#
*
#
*
!
(
(
!
! 24 Hour Counts- 1 Counter
#
(!(
!
(!
#
*
ON
(1 Bicycle)
(2 Bicycles)
(3 Bicycles)
(4 Bicycles)
(Motorcycle)
(Passenger Car)
(Car with Trailer)
(Rigid-body 2 Axle Truck)
(Rigid-body 3 Axle Truck)
(Rigid-body 4 Axle Truck)
(3 Axle Articulated Truck)
(4 Axle Articulated Truck)
(5 Axle Articulated Truck)
(6 Axle Articulated Truck)
(B-Double)
(Double/ Road Train)
1.056
1.827
1.657
NY
CA
1: 1Bike 2: 2Bikes
3: 3Bikes
4: 4Bikes
5: MC 6: SV 7: SVT 8: TB2 9: TB3
10: T4
11: ART3
12: ART4
13: ART5
14: ART6
15: BD
16: DRT (Bicycle)
(Motorcycle)
(Passenger Car)
(Car with Trailer)DRT non standard
(Rigid-body 2 Axle Truck)
(Rigid-body 3 Axle Truck)
(Rigid-body 4 Axle Truck)
(3 Axle Articulated Truck)
(4 Axle Articulated Truck)
(5 Axle Articulated Truck)
(6 Axle Articulated Truck)
(B-Double)
non standard
(Double/ Road BD
Train)
947
547
604
R
1: Cycle 2: MC 3: SV 4: SVT 5: TB2 6: TB3
7: T4
8: ART3
9: ART4
10: ART5
11: ART6
12: BD
13: DRT 2.68%
9.85%
8.95%
This table shows the accuracy and margins of error for counting bicyclists
using a MetroCount counter and our three ingredients for success.
ILE
RM
BOCO Vehicle Classes
94.68%
54.72%
60.35%
U
FO
ARX Cycle Vehicle Classes
1000
1000
1000
ED
For example, even though bicyclists’ axle
hits were now being recorded, the ARX
classification scheme was interpreting
two bicyclists to be a T4 (four axle) truck.
Multi-bike classes were necessary because once
the counters determine that a group of bicyclists
is one
vehicle, the classification scheme cannot
ART3
Non standard
split
the vehicle up. The new classes in the
BOCO scheme allow all bicyclists to be counted.
For example, a 2bike “vehicle” is actually two
bicyclists.
4'
27'
33'
HIN
NS
plugged
into a
single
counter
10.00%
ARX Cycle Rules for T4:
35
Distance
95%
Number Correction Corrected Margin
Number Counter
to
Confidence of
bikes
Factor
Number of Error
ART5
of Bikes Accuracy
Counter
Interval counted
(WB-62) (1/accuracy) of Bikes (bikes)
SU
anchored
with a
metal
bracket
20.00%
d3
LEE HILL D
two
tubes set
across
the road
30.00%
d2
d1
d1
N 45TH ST
the starting point:
40.00%
good
accuracy
when only
one person
is
riding in lap
s
N 63RD ST
0
bad accuracy when
when several people
are riding in laps
30
This graph shows the accuracy and margins of error for the Eco-Counter,
and the MetroCount counters, using both the ARX and BOCO classification
schemes.
T4 (non standard
OLDE STAGE RD
3-4’
0
Accuracy
4-6’
25
up
ro
1
20
'G
33
70.00%
accuracy when
d
o
o
g
60.00%
is
only one person
riding in laps
50.00%
15
Distance from counter (ft.)
ART5
Eco (WB-50)
BOCO 5' Loop
ARX 5' Loop
80.00%
2
10
up
ro
Creating the 2Bikes Class
5
up
ro
Editing the T4 Class
90.00%
3
Our solution was to create a new classification
scheme based on the ARX Cycle Scheme. We
titled it “Boulder County” (BOCO). We edited
the rules for the truck classes to exclude groups
of bicyclists. Then we created new multi-bike
classes to include groups of bicyclists.
60.00%
'G
27
100.00%
70.00%
N 95TH ST
Jay Rd: Number of Bicyclists Riding & Accuracy
TB2
80.00%
SVT
N 79TH ST
SV
N 75TH ST
bike
3. the new BOULDER COUNTY (BOCO) classification scheme
Staff and volunteers rode in laps across
the tubes to increase the amount of
data collected, but whenever more
than one person rode, we saw a large
drop in accuracy.
27 ft
The accuracy of the counters was dependent on the distance from the
counter bicyclists rode over the tubes. We split the data into three distance
groups: a 4 ft group, a 27 ft group, and a 33 ft group.
motorcycle
bike
4 ft
G
4'
Vehicle Wheelbase (ft)
5
33 ft
(
!
(
!
!
(
(
!
!
(!(
!
(
(
!
!
(
(
!
(
( (
!
!!
(!
(
!
(
!
BASELINE RD
N 76TH ST
6
motorcycle
4 ft
N 73RD ST
Speed vs. Wheelbase - Motorcycle Class
We discovered through
trial and error that
routing the “bicycle”
tubes through a vinyl
sleeve and securing
the sleeve with a metal
bracket secured the
tubes tightly without
pinching them.
Roads with Shoulders
N 71ST ST
the impetus:
thinner
“Bike”
tube
Roads without Shoulders
N 63RD ST
The purpose of our research study was to
determine if bicyclists could be counted
accurately using the county’s existing motor
vehicle traffic counters, which would provide
far better data about the county’s bicycle traffic.
standard
“Road”
tube
Thinner-walled
“bicycle”
tubes
allow bicyclists to
generate stronger
air pulses, which
have a greater
likelihood
of
being registered
as axle hits.
Results & Implementation:
Counter Accuracy
Prior to 2013, Boulder County has conducted
annual motor vehicle traffic counts at
approximately 175 locations using automated
pneumatic tube counters. The counters provide
detailed data including traffic volumes, speeds
and types, or classes, of vehicles on the county’s
road network. By contrast, bicycle counts were
limited to a handful of locations counted by
volunteers for just two hours each.
2. an attachment method that does not
pinch the tubes
N 61ST ST
1. thinner-walled “bicycle” tubes
Krista Nordback, Ph.D, PE
Researcher
Oregon Transportation Research
and Education Consortium (OTREC)
503.725.2897
[email protected]
CHERRYVALE RD
the goal:
We determined there are 3 necessary ingredients for counting
both bicyclists and motor vehicles at the same time, with the
same equipment:
SPINE RD
Research &
Development:
N 57TH ST
Bicycle Counts with
Pneumatic Tube Counters
Brian Graham, AICP
Bicycle Planner/
Employee Transportation Coordinator
Boulder County Transportation Department
720.564.2667
[email protected]
Alex Hyde-Wright
Transportation Engineering Assistant
Boulder County
Transportation Department
303.441.4910
[email protected]
(
!
(!
!!
(
(
(
!
In 2013, Boulder County is counting bicyclists and motor vehicles on all
paved county roads!
(
(!
!
(
!
Download this poster, our presentation/talking points and
paper here: http://goo.gl/UtDyB